For any naïve scientist finding a significant effect equates to proving a hypothesis. This is wrong. When conducting research, certain criterion is followed; stating your hypothesis, collecting and judging your data. There is a difference between deciding whether your conclusions are correct and proving your hypothesis. When conducting research, the scientific method should be used. This method uses both induction and deduction to assume that the findings from research will provide evidence for the desired effect. Rather than making individual assumptions, as this can lead to biases. Karl Popper was an influential figure in backing the use of the scientific method back in 1938. His assumption, which is still adhered to today was to generate a hypothesis, try to prove it wrong and then from your results generate a new hypothesis.
So why not try to prove your hypothesis? Simply because you can’t, it is virtually impossible to achieve as you would need to test 100% of the population, for a hypothesis to be proved correct. For example, it is common that we would all assume that apples are round, but is this always the case? Just because 99.9% of the world’s apples are round….
It doesn’t mean to say you won’t ever find a square one. Because it is impossible to prove a hypothesis, researchers tend to favour the process of falsification, outlined by Popper.
The idea that you can/should prove your hypothesis has a detrimental effect on science, as what is thought to be correct one minute, can change the next. We as scientists use our research to develop our understanding of the world us and make predictions about future event, or in psychology behaviours.
However, surly proving a hypothesis rather than falsifying it holds more weight? Well, if we are following the wise words of a famous proverb then perhaps proving something happened is easier than disproving the matter, if we don’t know it exists…. If a tree falls in the forest, and no one is around to hear it, does it make a sound?
In order to avoid personal biases occurring when conducting research, two hypotheses are generated. The experimental hypothesis which state the predicted direction of the results. And the null hypothesis which counteracts the biases of the experimental hypothesis as a prediction of no direction is made. It may simply be that no relationship will be found between two variables Null hypotheses are preferred as they are easily falsifiable in scientific research.
With this in mind the ultimate reason why a hypothesis cannot be proven is because it is a null hypothesis. A data set can only reject a null hypothesis, it can never be proven. For instance if 2 conditions are not statistically significant, it means that there is not enough evidence to reject the null….. details for why this occurs are saved for another time J
My Blog 